The Case of the Creeping BMI

By Ruth Kava — May 16, 2016
According to a new Danish study, obesity isn't as bad for health as it used to be. More exactly, the BMI associated with the lowest mortality risk seems to be higher than it was 40 years ago. But given some of the problems associated with using BMI to estimate obesity, we're not so sure that these results apply to everyone.
shutterstock_179390321 BMI categories via Shutterstock

Our faithful readers know (see here and here) that there are problems with using the body mass index (BMI : kgm/m2) as an indicator of obesity.

To recap quickly: BMI is most accurate when used to assess sedentary people — not  active folks, and certainly not elite athletes. The latter will have more muscle than fat, and muscle is heavier than fat, leading to spurious results. Further, while BMI when used appropriately can indicate excess body fat, it doesn't tell you where on the body the fat is located — and that can be important in assessing health risks. Finally, BMI is most useful and was designed for assessing the fatness and trends in fatness of populations, not individuals.

So what are we to make of a new report from Denmark, published in JAMA, that says that obesity isn't as much of a health risk as it used to be? Led by Dr. Shoalb Afzal from Copenhagen University and the Copenhagen University Hospital, a team of researchers analyzed data from three cohorts of the general Danish population. There were nearly 14,000 people from the Copenhagen City Heart Study (1976-78), another 9,500 from that study in 1991-94, and over 97,000 from the Copenhagen General Population Study in 2003-2013. All the participants were followed from their first enrollment in their respective study until November, 2014, or until they emigrated or died.

A so-called normal BMI is that associated with the lowest risk of death or disease, as indicated below, and is a J- or U-shaped curve.

 

Screen Shot 2016-05-16 at 10.57.23 AM BMI VS Mortality Risk via NIH

 

Typically the lowest risk of mortality is associated with a BMI in the lower to mid 20s. In the earliest Danish cohort, that nadir was 23.7. In the next one (1991-94) it was 24.6, and 27.0 for the last one. What could the progression in the nadir of the BMI VS mortality curve mean? Are the health effects of obesity and overweight really less than they were?

That could be the case — at least for Danes — whether it would also hold for other, more diverse populations isn't known. But could it also be the situation that Danes are exercising more and thus BMI is even less accurate than it was 40 years ago?

In either case, Dr. Gorge G. Nordestgaard commented: “Importantly, our results should not be interpreted as suggesting that now people can eat as much as they like, or that so-called normal weight individuals should eat more to become overweight." Nor, given the issues raised about BMI, should we simply assume these results will hold for all.